Estimating Track Mis-Registration Based on Readback Signal in Bit-Patterned Media Recording Systems Wiparat Busyatras * , Autthasith Arrayangkool * , Chanon Warisarn * , Lin M. M. Myint † , Pornchai Supnithi *‡ , and Piya Kovintavewat § * College of Data Storage Innovation, ‡ Telecommunications Engineering Department, King Mongkut’s Institute of Technology Ladkrabang 10520, Thailand. Email: s5690155@kmitl.ac.th, s4690151@kmitl.ac.th, kwchanon@kmitl.ac.th † School of Information Technology, Shinawatra University, Pathumthani 12160, Thailand. § Data Storage Technology Research Center, Nakhon Pathom Rajabhat University Nakhonpathom 73000, Thailand. Abstract—Track mis-registration (TMR) is one of crucial factors on the performance of bit-patterned media recording (BPMR) systems. Because of the movement of a read head and the rotation of a magnetic disk, the center of the read head may move away from its center location of a data track, thus resulting in a TMR effect, which can deteriorate the overall system performance. In general, the TMR effect can be detected and handled by a servo system. However, the servo system usually requires inherent sector-level latency mechanisms for head adjustment. Therefore, this paper proposes a method to estimate a TMR amount based only on a readback signal. Firstly, the signal-to-noise ratio (SNR) is estimated based on the peak amplitude of the readback signal. Then, we determine a TMR level using the average signal energy and the estimated SNR. Simulation results show that the proposed method can predict the TMR level embedded in the readback signal with 95% accuracy, especially when TMR and SNR level are high. Keywords—Bit-patterned media recording (BPMR); estimation method; signal-to-noise ratio (SNR); track mis-registration (TMR) I. I NTRODUCTION Track mis-registration (TMR) effect is a major obstacle for increasing an areal density (AD) in ultra-high density mag- netic recording systems such as bit-patterned media recording (BPMR) [1], [2]. The TMR effect is occurred due to the misalignment between the center of the read head and that of the main data track as illustrated in Fig. 1. Because of TMR, the readback signal may experience even more severe inter- track interference, which will further degrade the performance of data recovery process in BPMR systems [3], [4]. To handle the TMR effect, a servo system provides an inherent sector-level latency of the detection of servo bursts before any head adjustment can be made. This servo burst field has the information that can be used to estimate the amount of read head offset. However, it is generally difficult to predict the TMR quantity to the next servo sector, when TMR is compensated by only burst signals, especially when TMR goes beyond the limit [5], [6], [7]. In practice, several TMR estimation methods based on the readback signal have been proposed in the literature [8], [9]. Nonetheless, these methods require the knowledge of some Δ T x T z T Main Track Read Head Upper Track Lower Track Fig. 1. TMR in a BPMR system, i.e.,Δ T . recorded data and employ high complexity processing (e.g., calculating many correlation functions) for estimating TMR levels. Thus, this paper proposes a novel TMR estimation method, which is based only on the readback signal. Specifi- cally, at high SNR, our method can provide up to 95% accuracy at high TMR level (i.e., 20%-25%). This work begins with studying the relationship among the statistical information of the readback signal, signal-to- noise ratios (SNRs), and various TMR levels. Hence, these relationships will be employed to generate the mathematical equations to use estimating the SNR and the TMR level. In this work, we first estimate the SNR using the peak amplitude of the readback signal from one data sector. Then, this estimated SNR level will be utilized to estimate the TMR level according to the readback signal energy. This paper is organized as follows. In Section II, the BPMR channel model is described. Section III explains the proposed TMR estimation method. The simulation results are given in Section IV. Finally, Section V concludes this paper. II. CHANNEL MODEL In this paper, we focus on a discrete BPMR channel model [5], [6] as depicted in Fig. 2. The readback signal of the k th data bit on the main track can be expressed as r 0,k =  n  m h -m,n x -m,k-n + n 0,k , (1) The 29th International Technical Conference on Circuit/Systems Computers and Communications (ITC-CSCC), Phuket, Thailand, July 1-4, 2014 881